Re: [PATCH] minor doc fix for create-role

From: Tara Anne <tara(at)anne(dot)cat>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Re: [PATCH] minor doc fix for create-role
Date: 2019-08-19 01:00:46
Message-ID: 23A53972-985A-4E64-87CF-623D9EAFE51B@anne.cat
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi Tom,

I agree mostly. It actually does have the words “SQL identifier” in the patch. But you are right it doesn’t link to what a SQL identifier is, but it does provide a practical solution of quoting. That was the part I cared about as a user, I just wanted to solve my problem of an email address as a role name (yes I know that’s sort of dumb as email addresses change). This also addresses the question, why just here, because this was a pain point in the docs for me yesterday :)

I also agree your ideal solution is definitely better than what I pushed. But I’m not ready to take that on. If someone else is, I welcome their patch over mine.

-Tara


“Rivers know this: there is no hurry. We shall get there some day.”

> On Aug 18, 2019, at 9:41 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>
> tara(at)anne(dot)cat writes:
>> Attached is a minor patch to fix the name param documentation for create role, just adding a direct quote from user-manag.sgml talking about what the role name is allowed to be. I was searching for this information and figured the reference page should have it as well.
>
> Hm, I guess my reaction to this proposal is "why just here?". We have
> an awful lot of different CREATE commands, and none of them say more
> about the target name than this one does. (Not to mention ALTER, DROP,
> etc.) Perhaps it's worth adding some boilerplate text to all those
> places, but I'm dubious.
>
> Also, the specific text proposed for addition doesn't seem that helpful,
> since it doesn't define which characters are "special characters".
> I'd rather see something like "The name must be a valid SQL identifier
> as defined in <link to section 4.1.1>." But, while that would work fine
> in HTML output, it would not be particularly short or useful in man-page
> output.
>
> Perhaps the ideal solution would be further markup on the synopsis
> sections that somehow identifies each term as an "identifier" or
> other appropriate syntactic category, and provides a hyperlink to
> a definition (in output formats that are friendly to that). Seems
> like a lot of work though :-(
>
> regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Smith, Peter 2019-08-19 01:23:31 RE: [Proposal] Table-level Transparent Data Encryption (TDE) and Key Management Service (KMS)
Previous Message Tom Lane 2019-08-19 00:50:51 Re: Unused header file inclusion